Show Qwen Prompt

The "show me what the model said" node

Show Qwen Prompt is the third and smallest node in the ComfyUI-QwenVideo pack: a plain text viewer for the prompt that Qwen Video Prompt Reversal generates. You connect the main node's prompt STRING output to this node's text input, run the graph, and the generated description renders in the node's widget area instead of staying invisible in the console. That's it. No API calls, no model, no key - pure display.

Which is more useful than it sounds, because the main node already prints the generated prompt to the terminal as PROMPT PREVIEW, but most people don't watch the console. A display node puts the text where your eyes are, and for an API-backed tool you're paying per call, seeing the output before you feed it downstream is the difference between a reusable prompt and a wasted request.

The one useful trick: it's not a dead end

The detail that makes this node worth grabbing instead of grabbing any generic text-display node is that it's a pass-through, not a sink. Its schema is text in, text out - the same STRING comes back on the output, marked as a list. So you can hang it off the end of a chain for inspection and still keep the prompt flowing to the next node, or wire the output onward if you want to log the prompt at a specific point in a longer pipeline. It also handles list input gracefully (INPUT_IS_LIST is on in the source), so if a node upstream hands you a batch of strings it normalizes them to a list instead of throwing a shape error.

The node is marked as an output node and returns its text via the ui dict, which is how the text actually lands on the canvas.

Install

Same story as the rest of the pack - there's no way to get this node without the others:

cd ComfyUI/custom_nodes
git clone https://github.com/eddyhhlure1Eddy/ComfyUI-QwenVideo.git
cd ComfyUI-QwenVideo
pip install -r requirements.txt

Or search "QwenVideo" in ComfyUI Manager and restart. Dependencies are just requests, Pillow, torch, numpy; it needs no FFmpeg of its own (that's the main node's problem) and no API key.

The honest caveat

There's one thing the README dances around: "No text displayed in UI" is listed under troubleshooting, and the honest answer is that if you run a workflow and nothing appears on the node, the usual culprits are the graph not being refreshed or the node not being connected to an executed path - but also, on a fresh run, the display only fills in after the main node actually completes, which means waiting out that 15-25 second API call. Don't assume the node is broken; a blank text widget here usually means the upstream call failed first, and the real error is in the console. For what it does - show you the prompt, let it keep flowing - it does the job with zero fuss.